[sdiy] SAD1024 different modes

[Magnus Danielson]

From: "JH." <jhaible at debitel.net>
Subject: [sdiy] SAD1024 different modes
Date: Mon, 29 Aug 2005 18:04:29 +0200
Message-ID: <001101c5acb3$5af1b2c0$0200a8c0 at jhsilent>

> Hi,

Hi Jürgen,

> I was wondering about the usefullness of the different modes of operation
> for
> a SAD1024, in the case that I only need 512 stages.
> 
> I could
> 
> (a) just use one section
> 
> (b) connect both sections in parallel (as done in an Electro Harmonix
> Flanger)
> 
> (c) use both sections with differential amps (data sheet figure 10)
> 
> (d) use both sections in parallel/multiplex operation (data sheet figure 9)

I have not done anything with BBDs, so I have no real idea, but this made me
think. What about chopping a diffrential drive on both sides of the BBD. Thus,
for each sample the polarity is reversed and both stages would run inversed to
each other at all times. The chopping can be done by gating a signal of
opposite sign and double amplitude into a summing node at both sides.

This may have not additional benefit, but you could have some usefull averaging
for the lower end in terms of distorsion besides the benefit to be expected
from the balancing itself. It may just be a way to get higher complexity which
doesn't give any gain.

> The 5th method, serial connection and higer clocking, is not useful for my
> application
> (physical modeling), because I want the create very short delays, so the
> upper
> clock rate limit is a factor here.
> 
> I guess (a) has no benefits except less loading of the clock generator.
> 
> (b) looks trivial - might give 3dB SNR improovement over (a))
> 
> (c) is said to give cancellation of even order distortion

Of the a-c, c is my choice.

> (d) most interesing version - but is it the best solution?

Need to investigate this one.

Cheers,
Magnus